Merge pull request #32 from JuliaDecisionFocusedLearning/cleanup-argmax

Improve argmax benchmarks

Author: BatyLeo

Project.toml

@@ -1,7 +1,7 @@
 name = "DecisionFocusedLearningBenchmarks"
 uuid = "2fbe496a-299b-4c81-bab5-c44dfc55cf20"
 authors = ["Members of JuliaDecisionFocusedLearning"]
-version = "0.2.3"
+version = "0.2.4"
 
 [deps]
 Colors = "5ae59095-9a9b-59fe-a467-6f913c188581"

src/Argmax/Argmax.jl

@@ -40,6 +40,10 @@ function ArgmaxBenchmark(; instance_dim::Int=10, nb_features::Int=5, seed=nothin
     return ArgmaxBenchmark(instance_dim, nb_features, model)
 end
 
+function Utils.is_minimization_problem(::ArgmaxBenchmark)
+    return false
+end
+
 """
 $TYPEDSIGNATURES
 

src/Argmax2D/Argmax2D.jl

@@ -40,11 +40,15 @@ Custom constructor for [`Argmax2DBenchmark`](@ref).
 """
 function Argmax2DBenchmark(; nb_features::Int=5, seed=nothing, polytope_vertex_range=[6])
     Random.seed!(seed)
-    model = Chain(Dense(nb_features => 2; bias=false), vec)
+    model = Dense(nb_features => 2; bias=false)
     return Argmax2DBenchmark(nb_features, model, polytope_vertex_range)
 end
 
-maximizer(θ; instance) = instance[argmax(dot(θ, v) for v in instance)]
+function Utils.is_minimization_problem(::Argmax2DBenchmark)
+    return false
+end
+
+maximizer(θ; instance, kwargs...) = instance[argmax(dot(θ, v) for v in instance)]
 
 """
 $TYPEDSIGNATURES
@@ -56,7 +60,7 @@ function Utils.generate_dataset(
 )
     (; nb_features, encoder, polytope_vertex_range) = bench
     return map(1:dataset_size) do _
-        x = randn(rng, nb_features)
+        x = randn(rng, Float32, nb_features)
         θ_true = encoder(x)
         θ_true ./= 2 * norm(θ_true)
         instance = build_polytope(rand(rng, polytope_vertex_range); shift=rand(rng))
@@ -84,23 +88,30 @@ function Utils.generate_statistical_model(
 )
     Random.seed!(rng, seed)
     (; nb_features) = bench
-    model = Chain(Dense(nb_features => 2; bias=false), vec)
+    model = Dense(nb_features => 2; bias=false)
     return model
 end
 
+function Utils.plot_data(::Argmax2DBenchmark; instance, θ, kwargs...)
+    pl = init_plot()
+    plot_polytope!(pl, instance)
+    plot_objective!(pl, θ)
+    return plot_maximizer!(pl, θ, instance, maximizer)
+end
+
 """
 $TYPEDSIGNATURES
 
 Plot the data sample for the [`Argmax2DBenchmark`](@ref).
 """
 function Utils.plot_data(
-    ::Argmax2DBenchmark, sample::DataSample; θ_true=sample.θ_true, kwargs...
+    bench::Argmax2DBenchmark,
+    sample::DataSample;
+    instance=sample.instance,
+    θ=sample.θ_true,
+    kwargs...,
 )
-    (; instance) = sample
-    pl = init_plot()
-    plot_polytope!(pl, instance)
-    plot_objective!(pl, θ_true)
-    return plot_maximizer!(pl, θ_true, instance, maximizer)
+    return Utils.plot_data(bench; instance, θ, kwargs...)
 end
 
 export Argmax2DBenchmark

src/Argmax2D/polytope.jl

@@ -21,18 +21,16 @@ function plot_polytope!(pl, vertices)
         fillcolor=:gray,
         fillalpha=0.2,
         linecolor=:black,
-        label=L"\mathrm{conv}(\mathcal{V})",
+        label=L"\mathrm{conv}(\mathcal{Y}(x))",
     )
 end;
 
-const logocolors = Colors.JULIA_LOGO_COLORS
-
 function plot_objective!(pl, θ)
     Plots.plot!(
         pl,
         [0.0, θ[1]],
         [0.0, θ[2]];
-        color=logocolors.purple,
+        color=Colors.JULIA_LOGO_COLORS.purple,
         arrow=true,
         lw=2,
         label=nothing,
@@ -47,81 +45,9 @@ function plot_maximizer!(pl, θ, instance, maximizer)
         pl,
         [ŷ[1]],
         [ŷ[2]];
-        color=logocolors.red,
+        color=Colors.JULIA_LOGO_COLORS.red,
         markersize=9,
         markershape=:square,
         label=L"f(\theta)",
     )
 end;
-
-# function get_angle(v)
-#     @assert !(norm(v) ≈ 0)
-#     v = v ./ norm(v)
-#     if v[2] >= 0
-#         return acos(v[1])
-#     else
-#         return π + acos(-v[1])
-#     end
-# end;
-
-# function plot_distribution!(pl, probadist)
-#     A = probadist.atoms
-#     As = sort(A; by=get_angle)
-#     p = probadist.weights
-#     Plots.plot!(
-#         pl,
-#         vcat(map(first, As), first(As[1])),
-#         vcat(map(last, As), last(As[1]));
-#         fillrange=0,
-#         fillcolor=:blue,
-#         fillalpha=0.1,
-#         linestyle=:dash,
-#         linecolor=logocolors.blue,
-#         label=L"\mathrm{conv}(\hat{p}(\theta))",
-#     )
-#     return Plots.scatter!(
-#         pl,
-#         map(first, A),
-#         map(last, A);
-#         markersize=25 .* p .^ 0.5,
-#         markercolor=logocolors.blue,
-#         markerstrokewidth=0,
-#         markeralpha=0.4,
-#         label=L"\hat{p}(\theta)",
-#     )
-# end;
-
-# function plot_expectation!(pl, probadist)
-#     ŷΩ = compute_expectation(probadist)
-#     return scatter!(
-#         pl,
-#         [ŷΩ[1]],
-#         [ŷΩ[2]];
-#         color=logocolors.blue,
-#         markersize=6,
-#         markershape=:hexagon,
-#         label=L"\hat{f}(\theta)",
-#     )
-# end;
-
-# function compress_distribution!(
-#     probadist::FixedAtomsProbabilityDistribution{A,W}; atol=0
-# ) where {A,W}
-#     (; atoms, weights) = probadist
-#     to_delete = Int[]
-#     for i in length(probadist):-1:1
-#         ai = atoms[i]
-#         for j in 1:(i - 1)
-#             aj = atoms[j]
-#             if isapprox(ai, aj; atol=atol)
-#                 weights[j] += weights[i]
-#                 push!(to_delete, i)
-#                 break
-#             end
-#         end
-#     end
-#     sort!(to_delete)
-#     deleteat!(atoms, to_delete)
-#     deleteat!(weights, to_delete)
-#     return probadist
-# end;

test/argmax.jl

@@ -14,6 +14,9 @@
     model = generate_statistical_model(b)
     maximizer = generate_maximizer(b)
 
+    gap = compute_gap(b, dataset, model, maximizer)
+    @test gap >= 0
+
     for (i, sample) in enumerate(dataset)
         (; x, θ_true, y_true) = sample
         @test size(x) == (nb_features, instance_dim)

test/argmax_2d.jl

@@ -13,6 +13,9 @@
     model = generate_statistical_model(b)
     maximizer = generate_maximizer(b)
 
+    gap = compute_gap(b, dataset, model, maximizer)
+    @test gap >= 0
+
     # Test plot_data
     figure = plot_data(b, dataset[1])
     @test figure isa Plots.Plot